Abstract
In his new book (From empiricism to expressivism: Brandom reads Sellars, Harvard University Press, Cambridge, 2015), Brandom offers a new argument against the viability of Sellars’ scientific naturalism. Brandom attempts to show that if the Sellarsian it scientia mensura principle is understood as implying that manifest-image objects exist only if they are identical to scientific-image objects, it is undermined by the ‘Kant–Sellars’ thesis about identity which implies that manifest-image objects cannot be identical to scientific-image objects. This conclusion can be evaded by construing the relation between manifest and scientific objects as weaker than that of identity, namely as a relation between manifest-image functional roles and scientific-image realizers. But Brandom again argues that even this weaker construal of the scientia mensura thesis is in conflict with another Sellarsian argument, this time against phenomenalism. It will be argued that this is not so. I will, moreover, suggest that the ‘function-realizer’ construal of the manifest-scientific image distinction is indeed tenable—especially if the process of determining the scientific-image realizers of functional roles specified in manifest-image is understood as the culmination of a self-correcting dynamic and diachronic process of conceptual change. Finally, I will argue that while Brandom is right to point out that Sellars’ adherence to the scientia mensura principle is based on a ‘unity-of-science’ view, he is wrong to think that his argument for the contrary conclusion (the ‘disunity-of-science’ view) is successful, because Brandom’s argument does not automatically tell against a weaker ‘unity-of-science’ view according to which incommensurability of explanatory levels in science is a pragmatically indispensable yet in principle dispensable feature of empirical inquiry.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
Note, for example, that a thesis which accepts the significance of the disunity of explanatory levels (studied by the different special sciences) for reasons of practical utility, is fully compatible with the view to the effect that the ultimate raison d’être of the division of intellectual labour in empirical inquiry is the creation of a coherent (single, complex and stratified) picture of the world and our place in it.
As Brandom puts this point: “Part of taking an inference to be materially good is having a view about which possible additional collateral premises or auxiliary hypotheses would, and which would not, infirm it. Chestnut trees produce chestnuts—unless they are immature of blighted. Dry, well made matches strike—unless there is no oxygen (Brandom 2015, pp. 141–142). Notice, importantly, that this means that material inference is in general non-monotonic. That is, the inference from p to q may be materially good, even though the inference from p & r to q is not (Brandom 2015, pp. 163–165).
According to the scientia mensura principle the scientific image exercises final normative authority over what does and does not really exist. Brandom takes it that this means that descriptive terms from the manifest image refer to things specifiable in descriptive terms from the scientific image, if they refer at all; if some nonscientific descriptive term refers to anything real (rather than presenting a mere appearance), it is only because it corefers with some scientific descriptive term. But, coreference of terms is identity of objects. In this way, Brandom construes the claim of the scientia mensura principle as follows: to exist requires being identical to some object specifiable in the language of eventual natural science (Brandom 2015, p. 62). However, as we shall see (in Sect. 3), this construal of the relation between the descriptive and explanatory resources of the manifest and scientific image does not capture Sellars’ actual position here.
Recall here that Brandom construes the identity relation between the manifest and the scientific image as belonging in the realm of reference and not in that of sense; hence, he could equally object to a view according to which scientific-image explanations change the meaning, but retain the reference or ‘denotation’ of manifest-image terms. Yet, notice that, according to Sellars, the change of meaning in scientific-image explanations of manifest-image phenomena works by eliminating manifest-image objects, which means that the successor scientific-image concepts cannot possibly be understood as retaining ‘the same’ reference as their manifest-image ‘predecessors’.
Here one might wonder how can we reconcile the Sellarsian ‘Kant–Sellars’ thesis to the effect that empirical descriptions are modally involved (and the Aristotelian metaphysical framework of objects and properties with which this thesis is intimately bound up), with Sellars’ view that at the ideal end of empirical-scientific inquiry we ought to be committed to a (radically non-Aristotelian) naturalism of absolute processes that calls for a level of empirical description that is beneath any modal involvement. Indeed, as Brandom himself recognizes, Sellars does not deny the intelligibility-in-principle of purely descriptive discourse that contains no explicit modal vocabulary (Brandom 2015, p. 134). As Sellars himself puts it: “The idea that the world can, in principle, be so described that the description contains no modal expression is of a piece with the idea that the world can, in principle, be so described that the description contains no prescriptive expression. For what is being called to mind is the ideal of a statement of “every-thing that is the case” which, however, serves, through and through, only the purpose of stating what is the case. And it is a logical truth that such a description, however many modal expressions might properly be used in arriving at it, or in justifying it, or in showing the relevance of one of its components to another, could contain no modal expression” (Sellars 1957, p. 283). Now, one way in which one might try to reconcile the ‘Kant–Sellars’ thesis with Sellars’ radically ‘amodal’ ‘absolute process’ naturalism is to hold that, for Sellars, amodal descriptive discourse could be intelligible only as a totally unreflective and unselfconscious kind of discourse, which belongs to the stage of human language “when linguistic changes had causes, but not reasons, [before] man acquired the ability to reason about reasons” (Sellars 1957, p. 307). Indeed, this is precisely Brandom’s own way of making sense of the apparent conflict between the two above Sellarsian theses. On my reading, although Brandom’s proposal here is on to something important (namely, the fact that modal discourse is essentially pragmatic, and reflects the framework in which a representer can be properly critical towards its past, present and future representings) it is not ultimately satisfactory as an interpretation of Sellars. Pace Brandom, I take it that Sellars’ ‘purely descriptive’ language of ‘pure processes’ does not represent a regression to a stage where human language did not have a metalanguage at all (and hence, were completely unreflective and uncritical), but should be instead understood as having the status of a regulative ideal—i.e. as the culmination of a self-critical, self-correcting process of conceptual development. The regulative ideal in question points towards a kind of cognitive (and practical) ‘utopia’ in which the critical/reflective resources of the metalanguage (including modal, normative and explanatory discourse) would be rendered dispensable or optional. And this would be so just in case the regularities in behaviour which are implied by those critical/reflective resources were fully materially realized in the physical world and its relevant material mediums (in our case, the behaviour (skills, habits) of embodied human beings). Hence, far from implying an impoverishment of the critical/reflective resources of discourse and a regression to more primitive stages of human language, the ‘purely descriptive’ naturalistic ‘pure-process’ language, is actually the expression of what Sellars calls ‘the picture of language triumphant’ drawn in the heart of language militant’ (Sellars 1957, p. 307). And, in this context, the ‘Kant–Sellars’ thesis about modality can be understood as an indispensable means (‘in the heart of language militant’) for improving the descriptive and explanatory resources of language so as to approximate the above ideal.
Here it might be objected that the fact that the mode of existence of normatively-laden manifest-image phenomena (their ‘practical’/‘expressive’ reality) does not, strictly speaking, have explanatory purchase, entails the counterintuitive view that e.g. assertions such as “He did X because he wanted to do the right thing” or “She did it because it is the right thing to do” are, despite all appearances, never good explanations. But, the objection goes, can we really be that confused about the nature of explanation? In reply, we could say that there is a sense in which normative explanations are perfectly all right and a sense in which they are not. More specifically, to say that normative explanations are never good explanations is to make a point—from a transcendental point of view across different frameworks—about the manifest image framework (and its explanations) as a whole, not in it. Most significantly, it does not amount to a rejection of an empirical belief formulated within the framework. That is, it is not to say that the sentence “He did X because he wanted to do the right thing” expresses an empirical proposition, which, though widely believed by common sense, has been shown by science to be false. As long as we are in the manifest-image framework we evaluate statements about an individual and the desires, motives, thoughts, beliefs or reasons that find expression in his actions, in terms of criteria provided by this very framework. The manifest-image framework is a tightly interconnected whole and the attempt to replace it piecemeal by fragments of the scientific image is a cause of philosophical confusion. Thus, according to Sellars, to say that normative explanations are never adequate is an expression of a wholesale (not piecemeal) rejection of the manifest-image framework itself, in its descriptive and explanatory dimension, in favour of another built around different, though not unrelated, categories—namely, the ideal scientific image. It is also to say that science is making available a more adequate framework of entities, principles and laws (in the dimension of describing and explaining the world), which, in principle at least, could serve all the perceptual-inferential-practical functions of the framework that we employ in everyday life (Sellars 1997, §42; 1963c, p. 97). Again, this is not to say that there is good reason for adopting this envisaged new scientific framework in practice. Indeed there are sound methodological reasons for not teaching ourselves to respond to perceptible situations in terms of constructs in the language of the scientific-image (Sellars 1963c, p. 97).
Notice, for example, that if empirical generalizations are understood as having exclusively manifest observational content then the observed fact to the effect that e.g. gases do not obey the Boyle–Charles empirical law at very high pressures does not follow lawfully from past observational generalizations if we are restricted to the conceptual resources of the ‘manifest’ observational framework alone and can only be conceived as a brute anomalous fact whose behaviour obeys no law (Sellars 1976, §48–56). Hence, the empirical generalizations formulated within the observational framework, being strictly speaking non-lawlike (accidental), cannot even be inductively confirmed by their instances, i.e. by the particular observational facts that fall within their purview (since induction always involves going beyond one’s evidence from actual cases to counterfactual ones and accidental generalizations are precisely not counterfactually robust). By contrast, if we use the conceptual resources of the scientific postulational framework, the instability of the behaviour of gases in very high pressures virtually disappears and, given the theoretical framework of the kinetic theory of gases (with its unobservable entities, properties and the laws that relate them), the behaviour of gases is considered again normal, lawful and only to be expected. At the same time, given the theoretical framework of the kinetic theory of gases, we can explain why empirical generalizations couched in manifest-image terms (the Charles–Boyle empirical law) are violated: this violation is ultimately an appearance (it only seems to be a violation) of a deeper (lawful) reality, whose entities, properties and laws are in a position to account for the fact that it presents itself to us as appearance (i.e. as usually obeying and occasionally violating the Charles–Boyle law). More to the point, an analogous case can be envisaged as regards the empirical generalizations formulated in folk-psychological (i.e. manifest-image) terms and those formulated in terms of a postulational scientific micro-theory of psychological phenomena.
An interesting consequence of this line of thought is that the counterfactual robustness of subjunctive conditionals formulated in manifest-image terms would thereby be ultimately compromised. That is to say, in the above cases of observationally unpredicted variation, we would be unable to consistently apply (project) the manifest-image ‘lifeworld’ concepts in question to fresh, unobserved, cases (at least insofar as the latter involve cases in which the usual, expected behaviour of empirical objects—as delineated by these ‘lifeworld’ kind terms—is violated in an unpredictable manner). In other words, no rule in this case would be formulated to guide the concept’s non-random application to these, genuinely fresh, cases—i.e. there would be no distinction between what seems to be the right application and what actually is one. Accordingly, it can be argued that it is exactly in order to secure that minimal semantic distinction, necessary for the very stability (i.e. the possibility of inductive projectibility) of our empirical concepts that we have to view the postulational, scientific-image vocabulary as essentially involved in the determination of the very meaning of manifest-image ‘lifeworld’ terms.
Although, as we shall shortly see, I take it that Sellars would not accept this charge, it should be stressed that his view about what is involved in a change of concepts for explanatory reasons is so radical that it at least invites the above criticism. For example, according to Sellars, accepting a theory for explanatory reasons not only involves accepting a new conceptual framework, different from and not intertranslatable with the ‘old’ observational framework, but, even more radically, it is also the case that this new conceptual framework goes hand in hand with a new observation framework. This means that, in this case, our reliable dispositions to respond to external stimuli by issuing linguistic reports would be recalibrated so that, in response to the same kinds of stimuli, we would issue reports using the vocabulary of the new framework. This is what Sellars has in mind when he argues that what were traditionally called ‘correspondence rules’ should not be understood as defining new theoretical concepts in terms of old observational ones, but rather as redefining old observational terms by means of new theoretical concepts (Sellars 1961, §50). Moreover, although Sellars is silent on this point, it can be argued that, in this way, the new framework might enable us to devise new methods of detecting and measuring the previously unobservable entities and magnitudes posited by the theory, and thus adding terms referring to what was previously unobservable to our new observation framework (see also Roberts 2004, pp. 10–11). In any case, it is obvious that according to this picture, successful theoretical explanation involves the radical redescription or reconceptualization of the content of the explanandum itself, raising thereby Brandomian worries about whether what is actually explained in this process of conceptual change is the same with what was originally thought to be in need of explanation.
Recall that at the heart of Brandom’s argument lies the plausible observation that manifest-image kinds are identified and individuated functionally, by their relations to things of other such (manifest-image) functional kinds in complex systems articulated by social norms. If we combine this observation with the view that those functionally individuated manifest-image kinds comprise our very ‘lifeworld’—at the level of which the distinction between the normative and explanatory function of manifest-image kinds does not make clear sense-, we end up in a position very close to that attributed here to Brandom. Brandom does not explicitly make this latter claim, but I take it that something resembling that view is crucial for Brandom’s argument to be carried through successfully. And it is precisely this second claim that will be put into question in this section.
We can make essentially the same point by highlighting the fact that, for Sellars, empirical descriptions are internally connected to explanations, i.e. to certain counterfactually robust inferences which are necessary for licensing the applicability of an empirical description on the basis of another (and in this domain of discourse, the scientific image has the last word) while non-empirical descriptions are individuated on the basis of prescriptive and evaluative considerations (which are not inferior but just (irreducibly) different from explanatory considerations).
Indeed, in what follows, it will be suggested that these two functions of manifest-image terms are not only compatible with one another but are indeed ‘dialectically’ related in the sense that the discrepancy between them (i.e. between what ought to be the case and what is (explained to be) the case) is constitutive of the very identity (in a non-Brandomian sense) of a manifest-image term.
It must be noted that this regulative ideal operative in our descriptive and explanatory practices does not itself have a descriptive and explanatory function. Moreover, it could be argued—albeit not within the confines of the present paper—that (1) this regulative ideal can be understood as the demand for a systematic, unified and stratified body of empirical knowledge capable of yielding arbitrarily general pictures of the world and our place in it at an arbitrary level of specificity in detail, and that (2) its ultimate practical underpinnings have to do with our deep need to be liberated, at the individual as well as the collective level, from natural, biological, anthropological and social constraints or limitations, which result either out of pure ignorance about the structure of physical and human nature or from socially ‘non-optimal’ material distribution of already existing intellectual and material resources.
For example, it might be thought that ‘water’ is picked out as ‘whatever stuff causes certain effects’. In this case it would be a schematic (manifest-image) sortal which would really be a placeholder for less schematic (scientific-image) sortals. And the latter, less schematic scientific-image sortals, in turn, would be the expression of the normative ideals built into our manifest-image concept of ‘water’ -i.e. of what ideally ought to be the case for something to count as ‘water’ (that is to say, scientific-image sortals would be justified by their superior explanatory power in accounting for explanatory anomalies of manifest-image observable ‘water’ phenomena). Note that the sortals that identify and individuate intentional states and episodes are similarly schematic.
However, it must be noted that, for Sellars, this process of ‘preservation through transformation’ of the normative ideal of our manifest-image descriptive and explanatory practices makes sense only if the scientific image is in a position, in principle at least, to replace the manifest-image framework in all its roles, including, importantly, its role in conceptually responding to worldly stimulations and in guiding action directly in the conceptual terms of the framework. For example, Sellars explicitly states that “to say that the [manifest-image] framework is phenomenal in a quasi-Kantian sense ... is to say that science is making available a more adequate framework of entities which, in principle, at least, could serve all the functions, and in particular, the perceptual functions of the framework we actually employ in everyday life” (Sellars 1963c, p. 97). Hence, strictly speaking, this ‘preservation through transformation’ process is ultimately cashed out only in terms of what Sellars calls the ‘stereoscopic fusion’ of the manifest and the scientific image (where, interestingly, the very distinction between the manifest and the scientific image eventually loses its point).
Recall also here that, for Sellars, the distinction between manifest-image ‘lifeworld’ kind-terms with essentially observational/reporting uses and purely theoretical (‘successor’, scientific-image) terms without non-inferential reporting uses, is methodological (or, epistemological) rather than ontological. That is, the difference concerns how we know about something (by means of inference), not the kind of thing we know about. And from this it follows that the status of an object as theoretical (unobservable) or observable can change over time. For example, the existence of Mendelian genes was initially postulated as an explanation of observationally unpredicted variations in the hereditary characteristics of pea plants; yet, later, Mendelian genes became able to be observed on the basis of (theoretically-driven) predictions about where to look as well as the use of better microscopes. (Recall here that, according to Sellars, even intentional states such as thoughts, desires, intentions, beliefs etc. were also initially theoretical (unobservable) entities postulated to explain behavioural anomalies, and later became observational, i.e. obtained a reporting role.) In this way, we can see that, in Sellars, the appearance-reality distinction as well as the manifest-scientific-image distinction are essentially dialectical: scientific-image unobservable entities postulated to explain anomalies (observationally unpredicted variations—both about the ‘external’ and the ‘internal’ world) in the original manifest image can be subsequently incorporated to our perceptual and practical engagements with the world, and thus become features of a new conceptually transformed manifest image, enabling us again to be manifestly ‘open’—albeit in a new way—to the world and ourselves.
In order for a certain experiential event to be recognized as an anomaly in need of removal in practice as well as in theory (as opposed to something to which we have an attitude of indifference), it has to be placed within a normative and goal-directed context of significance, i.e. a context of (however schematic) already established practical and theoretical goals (‘oughts’), the divergence from which is to be noted and avoided, or at least, considered as a problem in need of a solution. Likewise, in order for any kind of differential response towards this recalcitrant experience to count as a cognitive-epistemic means for its explanatory accommodation, it has to be placed within a context in which certain (however schematic) norms of correct/adequate explanation are already operative, the divergence from which is to be noted and avoided, or at least, considered as an urgent problem in need of a solution. Note that if there were no discrepancy between the expected behaviour of a certain phenomenon (as determined by epistemic and practical norms) and its actual behaviour, we would not be in a position to even recognize a change in its behaviour as an anomaly, nor would we have any reason to create practical and epistemic tools in order to account for it.
This is because, for Sellars, talk about their generic similarity in functional role itself functions as a device that precisely abstracts from their specific differences—as the latter are expressed in their radically different specific content. This point is expressed in propositions such as “Euclidean triangularity and Riemannian triangularity are (both) varieties of triangularity”, “Classical negation and intuitionistic negation are varieties of negation” or “Newtonian (length, velocity, simultaneity) and relativistic (length, velocity, simultaneity) are varieties of (length, velocity, simultaneity)” (Sellars 1973). Note that in all the above cases we can say both that the concept in question has changed (in its specific content) and that it has remained the same (in its generic features).
As is obvious, the construal of scientific-image concepts as analogical counterparts of manifest-image concepts is intimately related to the fact that the former are generically similar to the latter while at the same time being specifically different from them. Notice, for example, that the Sellarsian notion of analogical model (on the basis of which we construct successor scientific-image concepts out of predecessor manifest-image concepts) and its use in the process of conceptual revision allows one to hold all the following views simultaneously: (a) that a change of our concepts is not arbitrary (since it is based on, up to a point ‘controlled’, analogical extensions from an already understood domain of ‘meanings’), (b) that this process of change leads to the formulation of concepts which are in a certain (generic) sense similar to the ones that are replaced (to the extent that certain analogies of the model on the basis of which the successor concept is constructed are essential in understanding the latter), and (c) that those very concepts described in (b) are in a certain (specific-contentual) sense clearly different from the ones that are replaced (to the extent that certain analogies of the model on the basis of which the successor concept is constructed are rejected—i.e. are thought of as not being part of the literal content of the latter). (For a related Sellarsian view regarding scientific change see Brown 1985.)
I do not wish to deny here that there is a difference between a stronger view according to which there is an in principle incommensurability of explanatory levels in science and a weaker view to the effect that the incommensurability in question is justified for practical reasons (but see n. 23). I am merely claiming that Brandom’s argument for the above strong version of incommensurability is compatible with (i.e. does not necessarily exclude) the weaker version (which, as will be argued, is Sellars’ own view).
For an analysis of this crucial notion of a special science’s ‘intended purposes’ as well as the connection of the latter with the counterfactual robustness and ultimate explanatory autonomy of a special science’s ceteris-paribus laws see Lange (2002).
It is interesting, in this connection, to note that even Sellars, who believes that special sciences can be explanatorily unified (at least in principle) in an ideal scientific image, does not think that this entails the unification of the sciences themselves, i.e. of their specific subject-matter; neither does it entail the unification of their theoretical principles and methods, at least to the extent to which the latter may be practically indispensable means for discovering the laws and principles that are obeyed by the entities and processes ‘disclosed’ from the theoretical standpoint of each special science (Sellars 1963a, p. 21).
Note that if this ‘pragmatic’ approach to concept-formation is correct, it blurs the distinction between logical and pragmatic incommensurability. Yet this view would be still distinguished from the strong ‘incommensurabilist’ version of the ‘disunity-of-science’ view as the latter was described in this section.
References
Brandom, R. (2015). From empiricism to expressivism: Brandom reads Sellars. Cambridge, MA: Harvard University Press.
Brown, H. I. (1985). Sellars and conceptual change. Synthese, 68, 275–307.
Christias, D. (2017). Does Brandom’s ‘Kant–Sellars thesis about modality’ undermine Sellars’ scientific naturalism? In D. Barnbaum & D. Pereplyotchik (Eds.), Sellars and contemporary philosophy (pp. 183–196). London: Routledge (forthcoming).
deVries, W. (2005). Wilfrid Sellars. Chesham: Acumen Publishing.
deVries, W. (2012). Ontology and the completeness of Sellars’s two images. Humana. Mente Journal of Philosophical Studies, 21, 1–18.
Dupré, J. (1993). The disorder of things: Metaphysical foundations of the disunity of science. Cambridge, MA: Harvard University Press.
Fodor, J. (1974). Special sciences (or: The disunity of science as a working hypothesis). Synthese, 28, 77–115.
Lange, M. (2002). Who’s afraid of ceteris paribus laws? Or: How I learned to stop worrying and love them. Erkenntnis, 57, 407–423.
Roberts, J. (2004, February 27). Sellars, Van Fraassen, layer cakes, and scientific realism. Presented to the Department of Philosophy: Florida State University.
Sellars, W. (1957). Counterfactuals, disposition and the causal modalities. In H. Feigl, M. Scriven, & G. Maxwell (Eds.), Minesota studies in the philosophy of science (Vol. II, pp. 225–308). Minneapolis, MN: University of Minesota Press.
Sellars, W. (1961). The language of theories. In H. Feigl & G. Maxwell (Eds.), Current issues in the philosophy of science (pp. 57–77). New York: Holt, Rinehart & Winston.
Sellars, W. (1963a). Philosophy and the scientific image of man. In Science, perception and reality. London: Routledge.
Sellars, W. (1963b). Theoretical explanation. In B. Baumrin (Ed.), Philosophy of science: The Delaware seminar (Vol. II, pp. 61–78). New York: Wiley.
Sellars, W. (1963c). Phenomenalism. In Science, perception and reality. London: Routledge & Kegan Paul.
Sellars, W. (1965). Scientific realism or irenic instrumentalism: A critique of Nagel and Feyerabend on theoretical explanation. In R. Cohen & M. Wartofski (Eds.), Boston studies in the philosophy of science (pp. 171–204). Atlantic Highlands, NJ: Humanities Press.
Sellars, W. (1967). Science and metaphysics: Variations on Kantian themes. Atascadero, CA: Ridgeview Publishing Company.
Sellars, W. (1973). Conceptual change. In P. Maynard & G. Pearce (Eds.), Conceptual change (pp. 77–93). Dordrecht: D. Reidel.
Sellars, W. (1976). Is scientific realism tenable? Proceedings of the Philosophy of Science Association, II, 307–334.
Sellars, W. (1997). Empiricism and the philosophy of mind. Cambridge, MA: Harvard University Press.
Wilson, M. (2008). Wandering significance. Oxford: Oxford University Press.
Wittgenstein, L. (1958). Philosophical investigations. Oxford: Blackwell.
Acknowledgements
An early version of this paper was presented in the ‘Sellars in a New Generation’ conference at Kent State University (2015). I would like to thank Bob Brandom, Preston Stovall and Carl Sachs for their useful comments. Thanks also to David Pereplyotchik and two anonymous reviewers for their useful comments and suggestions on previous versions of this paper.
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Christias, D. On the proper construal of the manifest-scientific image distinction: Brandom contra Sellars. Synthese 195, 1295–1320 (2018). https://doi.org/10.1007/s11229-016-1271-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11229-016-1271-1